Sea migration pattern of two sea trout (Salmo trutta ... · Zool. Fennici 39: 221–235 ISSN 0003-455X • Sea migration of two sea trout stocks 221 ... Sea migration pattern of two

ANN. ZOOL. FENNICI Vol. 39 • Sea migration of two sea trout stocks 221Ann. Zool. Fennici 39: 221–235 ISSN 0003-455XHelsinki 10 October 2002 © Finnish Zoological and Botanical Publishing Board 2002

Sea migration pattern of two sea trout (Salmo trutta) stocks released into the Gulf of Finland

Irma Kallio-Nyberg1, Ari Saura2 & Pekka Ahlfors2

1) Finnish Game and Fisheries Research Institute, Quark Fisheries Research Station, Korsholmanpuistikko 16, FIN-65100 Vaasa, Finland

2) Finnish Game and Fisheries Research Institute, Institute, P.O. Box 6, FIN-00721 Helsinki, Finland

Received 2 April 2001, accepted 23 October 2001

Kallio-Nyberg, I., Saura, A. & Ahlfors, P. 2002: Sea migration pattern of two sea trout (Salmo trutta) stocks released into the Gulf of Finland. — Ann. Zool. Fennici 39: 221–235.

The sea migration of two sea trout stocks was examined in a tagging and transplanta-tion experiment. The two different stocks, which originated from the rivers Isojoki (Gulf of Bothnia) and Ingarskilanjoki (Gulf of Finland), were released in 1994–1996 as smolts (3991 and 3996 smolts) at two nearby sites in the Gulf of Finland. The tag recovery data received 5–19 months after release (Sep.–Nov.) from the Gulf of Finland were analysed. The recovery rate was about 8%. Multi-way contingency analysis showed that the origin of the stock affected the spatial and temporal sea distribution. The majority (63%) of the Ingarskilanjoki sea trout, but a smaller proportion (49%) of the Isojoki sea trout, were caught in coastal waters, near the release site. The stock-specifi c spatial sea distributions were not affected by the gear types used. The genetic threats of stocking are discussed.

Introduction

The brown trout (Salmo trutta L.) shows varia-tion in its tendency to migrate. Both migratory and resident populations, genetically different from each other, can coexist in the same water system, or both sympatric forms, but with dif-ferent migratory behaviours, can belong to the same population (Jonsson 1982, 1985, Jonsson & Jonsson 1993, Hindar et al. 1991, Skaala & Neavdal 1989). Sea trout (Salmo trutta m.

trutta) is anadromous, that is, it migrates to salt water environment to feed and returns to streaming freshwater to spawn (L’Abee-Lund et al. 1989). The sea trout is caught mainly in coastal home waters, although some individuals may have travelled in the open sea 100–600 km away from their home river (Skrochowska 1969, Toivonen & Ikonen 1978, Berg & Berg 1987, Pratten & Shearer 1983). Both immature and mature brown trout tend to migrate annually between sea and river (Jonsson 1985, Sturlaugs-

222 Kallio-Nyberg et al. • ANN. ZOOL. FENNICI Vol. 39

son & Johannsson 1996). Sea trout leave the river for the fi rst time as 2–7-year-old smolts (Jonsson 1985, Johannsson & Einarsson 1993). They migrate and grow, on average, for 1–4 summers in the sea before maturing (L’Abee-Lund et al. 1989). The straying rate of sea trout varies depending on the stock traits (Berg and Berg 1987, Pratten & Shearer 1983, Johannsson & Einarsson 1993). Tagging and transplantation experiments show that the components of sea migration — distance and direction — are stock-specifi c traits of sea trout (Svärdson & Fag-erström 1982, Jonsson et al. 1994). The migra-tion patterns of brown trout are also largely controlled by environmental factors (Bohlin et al. 1993).

Two sea trout stocks, those of the Isojoki and the Ingarskilanjoki, are used in stockings in the Gulf of Finland. We examined stock-specifi c sea migration patterns by comparing the spatial and temporal marine distributions of tag recoveries of these stocks. Other traits of these stocks, e.g., growth rate, are compared in the work of A. Saura and P. Ahlfors (unpubl.). The Isojoki stock, which originates from the Gulf of Both-nia, was also largely used in the Gulf of Finland, in the 1980s and 1990s. The Ingarskilanjoki stock is a new breeding stock that originates from the Gulf of Finland.

The use of a domesticated stock in stockings in large sea areas increases the risk of loss of genetic variation; and moreover transplantation poses a threat to original stocks (Hansen & Loescheke 1994). There are numerous weak natural stocks in the rivers of the Gulf of Finland (Marttinen & Koljonen 1989, Saura 1998). The risk posed to natural stocks by introductions depends, among other things, on the release site and genetic variation of the hatchery fi sh. Differ-ent traits are required of stocks introduced into coastal waters for sea-ranching or into rivers for enhancement purposes. The aim of this work is to compare the sea migration behaviour of these stocks and to discuss on these stocks with management in view.

The tagged experimental smolt groups of both stocks were released at two sites near the estuary of the Vantaanjoki in 1994–1996. We examined the effect of stock origin and release site on the spatial and temporal sea distribution.

We tested the null hypothesis that spatial or temporal sea distribution is independent of stock and release site. If phenotypic variation occurred in the sea migration pattern between stocks, the variation must be genetic and stock-specifi c, because the prevailing conditions were the same for both experimental groups before the release. The migration pattern deduced from the tag recovery data depends on the distribution of the fi shery and gear types used. Consequently, we also examined the interrelationships between stocks, recovery sites or times and the gear types used in trout fi shery. The tag recovery data of the experimental groups were analysed by log-linear models.

Material and methods

Sea trout stocks

The sea trout stocks used in the experiment originated from two rivers, the Isojoki, which discharges into the Gulf of Bothnia, and the Ingarskilanjoki, which fl ows into the Gulf of Finland. The sea trout in the Ingarskilanjoki (Degerby å; Hurme 1970) can ascend the river for 13 km to the Myllypato dam, which has closed the upper parts of the river to anadromous trout since the 1930s (Hurme 1970, Marttinen & Koljonen 1989) (Fig. 1). Since 1998 there has been a fi sh way in the dam that functions when there is enough water in the river (A. Saura, unpubl.). In the lower part of the river, there are 11 rapids (area 890 m2), fi ve of which provide major habitats for juvenile trout (Marttinen & Koljonen 1989). In status, the Ingarskilanjoki natural sea trout stock is vulnerable (Koljonen & Kallio-Nyberg 1991). Despite stocking with reared parr and smolts, natural production has been very weak since fl ood prevention work was done in the river in 1989 (Saura 1998).

The Isojoki is 75 km long and has a total area of rapids in the main river and tributaries of 27 hectares. Characteristic of this river system are the number of small streams and brooks in it. Both sea trout and resident trout are present (Ahvonen et al. 1993). The natural sea trout stock of the Isojoki is endangered (Koljonen & Kallio-Nyberg 1991).

ANN. ZOOL. FENNICI Vol. 39 • Sea migration of two sea trout stocks 223

The breeding history of the Ingarskilanjoki trout is short. The fi rst wild juveniles were caught in the river for rearing in 1987 and 1988 (Saura 1998). The parr were reared as brood fi sh in a hatchery. The offspring of these fi sh were kept as brood fi sh of the second generation, which started to produce eggs for stocking purposes in 1997. Currently, all stocked Ingarskilanjoki sea trout are offspring of these second-generation brood fi sh. The stock has not been renewed with natural spawners. The Ingar-skilanjoki trout differs genetically from other trout stocks living in Finnish river systems enter-ing the Gulf of Finland (Koljonen 1989).

The Isojoki trout stock used in stockings originates from hatchery-reared brood stocks maintained in hatcheries for many fi sh genera-tions (Kallio 1986). It was the most widely used Baltic sea trout stock on the Finnish coast of the Gulf of Bothnia and in the Gulf of Finland in the 1980s and 1990s (Anon. 1992). Large anadromous sea trout spawners were favoured when the new brood stock was established with natural eggs. In the wild, local trout may spawn with migratory ones (Jonsson 1985), and in the Isojoki river system, too, non-migratory and migratory trout live partly in the same areas (Jutila et al. 1998). The Isojoki trout differs genetically from stocks living in Finnish rivers fl owing into the Gulf of Finland (Koljonen 1989). There are genetically different trout subpopula-tions in the Isojoki river system, and there are also clear genetic differences between the hatchery-reared brood stock and the natural parr sampled in the river (Ahvonen et al. 1993)

Experimental groups

The experimental groups were established with the offspring of spawners from the brood stock kept at the Laukaa hatchery station in central Finland. The groups released in 1994 were reared at this station, whereas those released in 1995 and 1996 were reared at the Savon Taimen hatchery station in the same region. All groups compared were raised separately under standard hatchery conditions for 2 years. The fi sh were held at Laukaa in plastic (fi rst year) and concrete (second year) tanks at ambient water tempera-

ture and were fed during daylight hours on com-mercial pelleted fi sh food. The rearing condi-tions were similar for both stocks.

One to six months before release, two-year-old pre-smolts were individually marked with external Carlin tags under MS 222 anaesthesia (Carlin 1969). The tag was attached by a double steel wire through the dorsal musculature of the fi sh. The weight of the tag (0.15 g) was 0.15% of that of the juvenile (100 g). The tagging was conducted in cold water (2–4 °C) in early spring, because the infection risk is then low and the fi sh has time to recover from the handling before release. Mortality during tagging was very low (under 0.5%). A total of 4000 juveniles of both stocks were tagged (Table 1). During tagging the total length (mm) of each fi sh was recorded; this was the length used as the smolt length of the group among recaptured fi sh.

The sea migration patterns of the two stocks were examined by comparing the tagged groups released in 1994–1996 at two sites near the estuary of the Vantaanjoki in the Gulf of Finland (Table 1). One release site was the mouth of the

��

��

� ��

��

��

��

��

�

��

� !�!�

!�"#��

$% $&

��'��

(��

)*�*� ��' ��*

+*��* ��

(�� ,*� ��

�

�

�

�

�%

Fig. 1. The Baltic Sea. The estuaries of the rivers Ingarskilanjoki, Vantaanjoki, Isojoki and Kyrönjoki are shown. Sea areas: Gulf of Finland, Gulf of Both-nia, Baltic main basin. Statistical squares 53 and 54 in the Gulf of Finland are drawn. The side of a square is about 55 km long.


Vantaanjoki river and the other was the shore of Hernesaari island. The distance between these two sites is about 8 km (Fig. 1). The groups compared were released every year at the same time in spring (April–May), when the water temperature is under 5 °C and the natural smolts also migrate into the sea.

Recovery data and fi shery

The distribution of recoveries collected was affected not only by the migration of trout but also by the distribution of fi shing, the catchabil-ity of fi sh and the gear types used in trout fi sh-ery. Trout were mainly caught as the by-catch of whitefi sh (Coregonus spp.) and pikeperch (Stizostedion lucioperca) in the Gulf of Finland

(Saura 1998). The distribution of recoveries therefore depends on the site and time of release. In salmon, the age and size of smolts also have an infl uence on migration behaviour (Hansen & Jonsson 1991a, Kallio-Nyberg et al. 1999).

Marking with external Carlin tags is a useful method in migration studies when comparing groups released at the same site or at the same time, assuming that the same variation exists between the groups in variables such as smolt size, which affect the behaviour of fi sh and their catchability (Berg & Berg 1987, Hansen et al. 1993, Hansen & Jonsson 1991a, 1991b). The visible tag is an important source of information on catch site and time. Many fi shermen also provide details of length, weight, sex and scale of fi sh and on the gear used in fi shing. There is some evidence that Carlin tagging has an

Table 1. Tagged experimental groups of Isojoki and Ingarskilanjoki stocks. Release year and site and number of fi sh tagged are shown.——————————————————————————————————————————————————————————————————Stock Year Release site No. of fi sh tagged——————————————————————————————————————————————————————————————————Ingarskilanjoki 1994 Vantaanjoki estuary 998Isojoki 1994 Vantaanjoki estuary 992Isojoki 1995 Vantaanjoki estuary 1000Ingarskilanjoki 1995 Vantaanjoki estuary 999Ingarskilanjoki 1995 Hernesaari 999Isojoki 1995 Hernesaari 999Isojoki 1996 Hernesaari 1000Ingarskilanjoki 1996 Hernesaari 1000——————————————————————————————————————————————————————————————————

Table 2. Number of tag recoveries and tag recovery rate (%) of experimental groups of Isojoki and Ingarskilanjoki trout stocks. Period: 1 = May–Aug. in release year, 2 = Sep.–Nov., in release year, 3 = Dec.–Aug. from fi rst to second year, 4 = Sep.–Nov. in second year, 5 = Dec.–Aug. from second to third year, 6 = Sep.–Nov. in third year, 7 = Dec.–Aug. from third to fourth year, 8 = Sep.–Nov. in fourth year, 9 = Dec.–Aug. from fourth to fi fth year. Annual spawning migration of trout takes place in September–November. Release site: estuary of the Vantaanjoki (Va) and shore of Hernesaari (He).——————————————————————————————————————————————————————————————————Stock Year/site Number of tag recoveries by periods Total Recov. number rate 1 2 3 4 5 6 7 8 9 ——————————————————————————————————————————————————————————————————Ing. 94/Va 16 8 16 14 6 0 1 0 0 61 6.1Iso. 94/Va 13 10 26 8 11 0 0 0 1 69 6.9Iso. 95/Va 20 6 14 4 9 1 0 1 0 55 5.5Ing. 95/Va 14 6 18 22 8 1 1 0 0 70 7.0Ing. 95/He 2 16 39 64 17 5 2 1 0 146 14.6Iso. 95/He 5 10 28 12 24 7 7 1 1 95 9.5Iso. 96/He 3 23 36 10 7 3 0 0 0 82 8.2Ing. 96/He 2 21 19 19 8 0 0 0 0 69 6.9——————————————————————————————————————————————————————————————————


adverse effect on the survival of tagged fi sh in the sea (Hansen 1988). The effect of tagging on the sea migration of fi sh is unknown, but comparative studies similar to the present one suggest that the differences in migration behav-iour between groups are signifi cant (Svärdson & Fagerström 1982). Carlin tagging is the only method able to give information on the migra-tion of individual fi sh.

In our analysis we deal with the Gulf of Finland, because 89% of the Isojoki sea trout and 93% of the Ingarskilanjoki sea trout were caught there. Within the Gulf of Finland the majority of the Isojoki and Ingarskilanjoki sea trout, 51% and 61%, respectively, were caught in coastal waters. These coastal waters were in statistical squares 53 and 54, which are located on the northern coast of the Gulf of Finland, near the release sites (Fig. 1).

The recovery time was divided into nine peri-ods according to seasonal migration (Table 2). In the fi rst months, 1–4 months after release, the post-smolts were not completely recruited to the fi shery. The second period was the fi rst spawning migration (5–7 months after release); the third period was the feeding migration (8–16 months in the sea) and the fourth period was the second spawning migration (17–19 months in the sea). In order to establish whether there was any difference in migration pattern between these two stocks, only recoveries made during 5–19 months in the sea were included in the analysis. Most of the fi sh were captured during these periods (Table 2).

In salmon (Salmo salar L.), the initial smolt size has an impact on migration pattern (Salminen et al. 1994, Kallio-Nyberg et al. 1999) and on catchability and fi shing mortality (Salminen

et al. 1995). In our data, during the period 5–19 months after release there was no signifi -cant difference in mean fi sh weight between these two stocks. The Isojoki sea trout showed, however, a faster increase in weight than did the Ingarskilanjoki sea trout in the second sea year (Fig. 2). In addition, no differences in mean smolt length were found between the stocks compared (t-test: t = 1.91, P = 0.056; Isojoki stock: 22.6 ± 2.08 cm, N = 301 and Ingarskilan-joki stock: 22.3 ± 1.94 cm, N = 347). The smolt lengths of sea trout were similar in the two experimental groups in both 1994 and 1996 (Table 3), as calculated from recovered fi sh. In 1995, however, the mean smolt length of the Iso-joki sea trout released on the shore of Hernesaari was greater than that of the Ingarskilanjoki trout in the same year (t-test: N = 95, N = 69, t = –3.43, P < 0.001) (Table 3). The variation in smolt size within a year and between years was not included in the analysis, because the groups compared did not show any differences.

�

��

��

��

��

��

��

��

��

��

��

� � � � ��

��

��

��

��

Fig. 2. Mean fi sh weight (± SD) in Ingarskilanjoki and Isojoki sea trout 5–19 months after release.

Table 3. Mean smolt size of experimental groups released in 1994–1996.

Stock Year/site Smolt length, (cm) N

Mean SD

Ing. 94/Va 23.3 2.07 61Iso. 94/Va 23.4 1.99 68Iso. 95/Va 20.3 1.79 56Ing. 95/Va 20.7 1.44 70Ing. 95/He 22.6 1.71 147Iso. 95/He 23.4 1.74 95Iso. 96/He 22.8 1.46 82Ing. 96/He 22.6 1.71 69


The mean lengths of the Isojoki and Ingarski-lanjoki sea trout were the same when calculated from fi sh captured during the fi rst sea winter (Dec.–Mar., from the fi rst to the second year) in the Gulf of Finland (Isojoki: 43.3 ± 5.93 cm, N = 47; Ingarskilanjoki: 41.7 ± 4.71 cm, N = 30; t = 1.22, df = 75, P = 0.225). However, after two growth periods in the second sea winter, the Isojoki trout were longer than the Ingarskilanjoki trout in the Gulf of Finland (Iso-joki: 63.0 ± 6.07 cm, N = 17; Ingarskilanjoki: 53.8 ± 7.36 cm, N = 8; t = 3.30, df = 23, P = 0.003). The corresponding differences in mean length between stocks were shown by fi sh caught in the Baltic Sea, including all recoveries in the fi rst or second sea winters.

The interactions between gear type and stock or between gear type and recovery site or recov-ery time were examined. However, not all the recovery reports included information on gear. Such information was available in 453 cases. A total of 25 gear types were used in trout fi shery. These were classifi ed into six groups: (1) fl oating gill net, (2) drift net, (3) trap net, (4) long line, (5) rods and (6) others. The fi rst class included all kinds of bottom gill nets, and bar lengths varied from 27 to 65 mm. In the second class the bar length was 80 mm, which is the smallest bar size allowed for fl oating gill nets in the Gulf of Finland.

Statistical analysis

The recovery data are presented in multi-way contingency tables and were analysed by log-

linear models (McCullagh & Nelder 1989) using the CATMOD procedure of the SAS statistical package (Anon. 1989). This method is well suited for analysing frequency data with several dimensions. The dependence of stock, release site, recovery site and time on both gear types was analysed by three-dimensional log-linear models. In the fi rst analyses, we tested the interrelationships between stock, recovery site and recovery time. The details of the analyses of migration pattern by log-linear models have been given by Kallio-Nyberg et al. (2000).

Results

Interrelationships between stock and recovery site and time

The majority of the sea trout migrated to the sea to feed near the release site and stayed there during four sea years. Eighty-nine per cent (N

all = 268) of the Isojoki and 83% (N

all = 314)

of the Ingarskilanjoki sea trout were recovered within the Gulf of Finland and only 11% and 7%, respectively, outside the Gulf, in the Baltic main basin or in the Gulf of Bothnia. About half of the Isojoki (51%, N

all = 259) sea trout were

recovered east of the release site (25°E). Cor-respondingly, 54% of the Ingarskilanjoki trout were caught in the eastern part of the Gulf of Finland and 46% (N

all = 306) elsewhere in the

Baltic Sea during three sea years (December in the third sea year).

Within the Gulf of Finland about half (49%) of the Isojoki sea trout and the majority (63%)

Table 4. Number of recoveries in three-dimensional contingency table. Variables: stock, recovery site and recovery time. Coastal squares = squares 53 and 54 in Gulf of Finland. Sea = Gulf of Finland, except squares 53 and 54, near the release site.——————————————————————————————————————————————————————————————————Recovery site Recovery time (months after release)

Isojoki stock Ingarskilanjoki stock

5–7 8–16 17–19 % 5–7 8–16 17–19 %——————————————————————————————————————————————————————————————————Coastal squares 24 38 12 49 28 43 68 63Sea 13 51 12 51 13 39 31 37——————————————————————————————————————————————————————————————————Percentage 25 59 16 18 37 45——————————————————————————————————————————————————————————————————


of the Ingarskilanjoki sea trout were caught in coastal waters within 5–19 months of their release (Table 4). There was some seasonal migration between coastal and offshore waters. In autumn (Sep.–Nov.: here 5–7 and 17–19 months after release) the trout migrated most frequently in coastal waters. From 68% to 69% of the Ingarskilanjoki trout and from 65% to 50% of the Isojoki trout were caught in these autumn months in coastal waters near the release site. During their feeding migration, 8–16 months after release, the Isojoki sea trout in particular migrated in the open sea and also in the coastal waters of Estonia. Of the Ingarskilanjoki trout 2.5%, and of the Isojoki trout over 6%, were caught more than 200 km from the release site. The most distant recaptures of the Isojoki trout

occurred more than 800 km from the release site, near Bornholm, an island in the southern Baltic Sea (Fig. 3).

The Isojoki and Ingarskilanjoki trout were recovered during different sea periods. The majority (59%, 89/150) of the Isojoki sea trout, but a smaller proportion (37%, 82/222) of the Ingarskilanjoki fi sh, were caught in winter and summer (Dec.–Aug.) during their feeding migra-tion, 8–16 months after release (Table 4). The Ingarskilanjoki trout were caught mainly in autumn during the spawning migration (Sep.–Nov.).

We analysed the interrelationships between stock, recovery site and recovery time by log-linear models. The simplest log-linear model with P > 0.05 featured two interactions between

Fig. 3. All recoveries of Isojoki and Ingarskilanjoki sea trout released into the estuary of the Vantaan-joki in 1994–1996.


stock and recovery time and recovery time and site (Model, H

0: S, T, P, S ¥ T, T ¥ P: G2 = 4.60,

P = 0.20; see Table 5). The model with three interac-tions was not signifi cantly better than this model with two interactions (G2

ST, TP – G2

ST, TP, SP = 3.67,

df = 3 – 2 = 1, the critical value of the h2 distribution with df = 1 and at the risk level 0.05 is 3.84). This shows a statistically signifi cant difference in temporal sea distribution between the two stocks. As pointed out above, the Isojoki sea trout were caught mainly in winter–summer and the Ingarskilanjoki trout in autumn. The interaction between recovery time and recovery site shows that either trout move seasonally between coastal waters and the open sea or that fi shing activity varies seasonally between sea areas. The majority (51%) of the Isojoki trout, but a smaller proportion (37%) of the Ingarskilanjoki trout, were caught in the open sea in the Gulf of Finland. This difference in spatial sea distribution between stocks was not, however, statistically signifi cant. The effect of release site was not included in our analysis.

Stock, release site and recovery site

The genetic characteristics of the stock, but not the release site, had an effect on the spatial sea distribution of fi sh. The one-interaction model with the interaction between stock and recovery site was compatible with our data (H

0: S, R, P, S ¥ P: P = 0.5), and was signifi -

cantly better than the model without the interac-tion (G2

S, R, P – G2

S, R, P, S ¥ P = 8.80 – 2.36 = 6.44;

df = 4 – 3 = 1; h20.05

= 3.84) (Table 6). The more complex model, with interactions between stock and recovery site and between release and recovery site, was even more compatible with the data (P = 0.551), but it was not sig-nifi cantly better than the one-interaction model (G2

S, R, P, S ¥ P – G2

S, R, P, S ¥ P, S ¥ R = 1.17, df = 3 – 2 =1;

h20.05

= 3.84). The release site was an independ-ent variable in the model used. The model that included the release site showed the effect of origin of stock on the spatial sea distribution. Forty-nine per cent (N

all = 150) of the Isojoki

trout and 63% (Nall

= 222) of the Ingarskilanjoki trout were caught in the coastal squares (Fig. 4).

Table 5. Effect of stock on spatial and temporal sea distribution in the Gulf of Finland. Log-linear models and test of independence between classifi ed varia-bles: stock, S (Isojoki and Ingarskilanjoki trout), recov-ery site; P (statistical squares 53 and 54 = coastal and other waters of the Gulf of Finland), recovery time, T (5–7, 8–16, 17–19 months in the sea). The contingency table with sample sizes is shown in Table 4. Variables are expected to be independ-ent when separated by a comma and to interact when there is no comma. Model, null hypothesis; df = degrees of freedom; G2 = likelihood ratio; °P-val-ues indicate the discrepancy between the model and the data; f* shows the models with the best fi t (P > 0.05).

Model df G2 °P

S, T, P 7 52.9 0.000S, T, P, S ¥ P 6 46.4 0.000S, T, P, S ¥ T 5 17.3 0.003S, T, P, P ¥ T 5 40.1 0.000S, T, P, S ¥ P, S ¥ T 4 10.9 0.029S, T, P, S ¥ P, T ¥ P 4 33.7 0.000S, T, P, S ¥ T, T ¥ xP 3 4.6 0.204f*S, T, P, S ¥ P, S ¥ T, T ¥ P 2 0.9 0.626f*

Table 6. Effect of stock and release site on spatial sea distribution in the Gulf of Finland. Log-linear models and test of independence between classifi ed variables: stock, S (Isojoki and Ingarskilanjoki trout), release site, R (estuary of the Vantaanjoki and coastal waters of Hernesaari island) recovery site, P (statisti-cal squares 53 and 54 = coastal and other waters of the Gulf of Finland). Sample sizes are shown in Fig. 4. Variables are expected to be independent when separated by a comma and to interact when there is no comma. Model, null hypothesis; df = degrees of freedom; G2 = likelihood ratio; °P-values indicate the discrepancy between the model and the data; f* shows the models with the best fi t (P > 0.05).

Model df G2 °P

S, R, P 4 8.80 0.066S, R, P, S ¥ P 3 2.36 0.500f*S, R, P, R ¥ P 3 7.63 0.054S, R, P, S ¥ R 3 7.64 0.054S, R, P, S ¥ P, R ¥ P 2 1.19 0.551f*S, R, P, S ¥ P, S ¥ R 2 1.20 0.550S, R, P, R ¥ P, S ¥ R 2 6.46 0.039S, R, P, S ¥ P, R ¥ P, S ¥ R 1 0.30 0.586


Stock, release site and recovery time

The origin of the stock, but not the release site, affected the recovery time. The one-interaction model between stock and recovery time was sig-nifi cantly better than the model without the inter-action (G2

S, R, T – G2

S, R, T, S ¥ T = 47.09 – 3.64 = 43.09;

df = 7 – 5 = 2; h20.01

= 9.21) (Table 7). The two-interaction model between stock and recovery time and between release site and recovery time was more compatible with our data (P = 0.902), but it was not statistically sig-nifi cantly better than the one-interaction model (G2

S, R, T, S ¥ T – G2

S, R, T, S ¥ T, R ¥ T = 3.64 – 0.57 = 3.24;

df = 5 – 3 = 2; h20.05

= 5.99). Fifty-seven per cent of the Isojoki trout (N

all = 169) and 34%

(Nall

= 248) of the Ingarskilanjoki trout were caught in the feeding period (Dec.–Aug.) (Fig. 5).

Stock, gear type and recovery site

The model with two interactions, between stock and recovery site and between gear type and recovery site, was compatible with our data (P = 0.412). It was signifi cantly better than the model with one interaction, between gear type

and recovery site (G2S, G, P G ¥ P

– G2S, G, P, S ¥ P, G ¥ P

= 13.39 – 8.22 = 5.17; df = 9 – 8 = 1; h2

0.05 = 3.84)

(Table 8). Adding the interaction between stock and gear type did not improve the model signifi -

�

��

��

��

��

��

��

��

�� !� "#$�#%��$�

&

� ' �� (� ��

Fig. 4. Spatial sea distribution of recoveries in the Gulf of Finland. Recovery site (two classes): Other areas (= Gulf of Finland, except squares 53 and 54, near the release site) and coastal squares (= squares 53 and 54). The release sites (estuary of the Vantaanjoki and shore of Hernesaari island) are located in square 53. Recoveries 5–19 months after release are included.

Table 7. Effect of stock and release site on temporal sea distribution in the Gulf of Finland. Log-linear models and test of independence between classifi ed variables: stock, S (Isojoki and Ingarskilanjoki trout), release site, R (estuary of the Vantaanjoki and shore of Hernesaari island) recovery time, T (5–7, 8–16, 17–19 months in the sea). The sample sizes are shown in Fig. 5. Variables are expected to be inde-pendent when separated by a comma and to interact when there is no comma. Model, null hypothesis; df = degrees of freedom; G2 = likelihood ratio; °P-val-ues indicate the discrepancy between the model and the data; f* shows the models with the best fi t (P > 0.05).

Model df G2 °P

S, R, T 7 47.09 0.000S, R, T, S ¥ T 5 3.64 0.603f*S, R, T, R ¥ T 5 44.03 0.000S, R, T, S ¥ R 6 46.32 0.000S, R, T, S ¥ T, R ¥ T 3 0.57 0.902f*S, R, T, S ¥ T, S ¥ R 4 2.87 0.579S, R, T, R ¥ T, S ¥ R 4 43.26 0.000S, R, T, S ¥ T, R ¥ T, S ¥ R 2 0.31 0.854


Fig. 5. Temporal sea dis-tribution of recoveries in the Gulf of Finland. The release sites (estuary of the Van-taanjoki and shore of Her-nesaari island) are located in square 53. Recovery time: (three classes): 5–7, 8–16 and 17–19 months after release.

�

��

��

��

��

��

��

��

��

��

� � ��

�� !"� #$%�$&��%�

'

�� (

Table 8. Interrelationships between stock, gear type and release site in the Gulf of Finland. Log-linear models and test of independence between classifi ed variables: stock, S (Isojoki and Ingarskilanjoki trout), gear type, G (classes 1–5: bottom gill net, fl oating gill net, trap net, long lines, rods) recovery site, P (statistical squares 53 and 54 = coastal and other waters of the Gulf of Finland. Recoveries 5–19 months after release are included. The contingency table with sample sizes is shown. Variables are expected to be independent when separated by a comma and to interact when there is no comma. Model, null hypothesis; df = degrees of freedom; G2 = likelihood ratio; °P-values indicate the discrepancy between the model and the data; f* shows the models with the best fi t (P > 0.05). ——————————————————————————————————————————————————————————————————Gear types Isojoki stock Ingarskilanjoki stock

Recovery site Total (%) Recovery site Total (%)

Coast Other areas Coast Other areas——————————————————————————————————————————————————————————————————Bottom gill net 50 37 67 77 46 59Floating gill net 4 1 4 19 1 10Trap net 2 12 11 7 11 9Long line 1 1 2 1 1 1Rods 8 13 16 27 18 21——————————————————————————————————————————————————————————————————Model df G2 °P——————————————————————————————————————————————————————————————————S, G, P 13 40.96 0.000S, G, P, S ¥ P 12 35.79 0.000S, G, P, G ¥ P 9 13.39 0.145S, G, P, S ¥ G 9 34.39 0.000S, G, P, S ¥ P, G ¥ P 8 8.22 0.412f*S, G, P, S ¥ P, S ¥ G 8 29.91 0.000S, G, P, G ¥ P, S ¥ G 5 6.81 0.235S, G, P, S ¥ P, G ¥ P, S ¥ G 4 2.98 0.561——————————————————————————————————————————————————————————————————


cantly (G2S, G, P, S ¥ P, G ¥ P

– G2S, G, P, S ¥ P, G ¥ P, S ¥ G

= 8.22 – 2.98 = 5.24; df = 8 – 4 = 4; h2

0.05 = 9.48).

Thus the origin of the stock had an effect on the spatial sea distribution in the two different sea areas where different gear types were used. The analysis also showed that stocks were not caught with different gear types. The stock-specifi c migration patterns were not due to the different gear types used. The majority of the Isojoki (67%) and Ingarskilanjoki (59%) sea trout were caught with gill nets (Table 8).

Stock, gear types and recovery time

The model with two interactions, between stock and recovery time and between gear type and recovery time, was most compatible with the data (P = 0.555) (Table 9) and was signifi cantly better than the model with one interaction (G2

S, G, P, S ¥ T – G2

S, G, P, S ¥ T, G ¥ T = 26.73;

df = 10 – 6 = 4; h20.001

= 18.46 and G2

S, G, P, G ¥ T – G2

S, G, P, S ¥ T, G ¥ T = 24.34; df = 8 – 6 = 2;

h20.001

= 13.81). This implies that the recovery time depended on the stock and that different gear types were used seasonally. The analysis also shows that both the Isojoki and the Ingarski-lanjoki trout were caught equally frequently with these three gear types. The majority of fi sh, 72% of Isojoki and 66% of Ingarskilanjoki trout, were caught with bottom gill nets, when three gear types were included in the analyses (Table 9). The bottom gill net catch was overwhelmingly domi-nant in the feeding period, December–August.

Discussion

Stock-specifi c sea migration

The two sea trout (Salmo trutta) stocks com-pared here differed signifi cantly from each other

Table 9. Interrelationships between stock, gear type and release time in the Gulf of Finland. Log-linear models and test of independence between classifi ed variables: stock, S (Isojoki and Ingarskilanjoki trout), gear type, G (classes 1, 3, 5: bottom gill net, trap net, rods) recovery time, P (5–7, 8–16, 17–19 months after release. Recoveries from the Gulf of Finland are included. The contingency table with sample sizes is shown. Variables are expected to be independent when separated by a comma and to interact when there is no comma. Model, null hypothesis; df = degrees of freedom; G2 = likelihood ratio; °P-values indicate the discrepancy between the model and the data; f* shows the models with the best fi t (P > 0.05).—————————————————————————————————————————————————————————————————— Isojoki stock Ingarskilanjoki stock

Recovery time Recovery (%) Recovery time Recovery (%)

Gear types 5–7 8–16 17–19 5–7 8–16 7–19——————————————————————————————————————————————————————————————————Bottom gill net 26 58 6 72 31 60 43 67Trap net 2 6 6 11 1 9 13 11Rods 7 8 6 17 11 11 23 22——————————————————————————————————————————————————————————————————Model df G2 oP——————————————————————————————————————————————————————————————————S, G, T 12 55.98 0.000S, G, T, S ¥ T 10 31.64 0.000S, G, T, S ¥ G 10 54.44 0.000S, G, T, G ¥ T 8 29.25 0.000S, G, T, S ¥ T, S ¥ G 8 30.10 0.000S, G, T, S ¥ T, G ¥ T 6 4.91 0.555f*S, G, T, G ¥ T, S ¥ G 6 27.72 0.000S, G, T, S ¥ T, S ¥ G, G ¥ T 4 4.19 0.384——————————————————————————————————————————————————————————————————


in spatial and temporal sea distributions. The Ingarskilanjoki trout migrated more frequently to feed in coastal waters, in the archipelago area, near the release site than did the Isojoki trout. The null hypothesis, that spatial sea distri-bution is independent of stock, was thus rejected. The stock-specifi c differences were expected, because the salmon stocks originated from dif-ferent parts of the Baltic Sea. The Gulf of Bothnia, which is the natural migration area for the Isojoki sea trout, and the Gulf of Fin-land, which is the home area of the Ingarskilan-joki sea trout, have different food resources for salmon (Salminen et al. 1995). The stocked Neva salmon, which migrates only a short dis-tance, grows more slowly in the Gulf of Bothnia than in the Gulf of Finland. Moreover, most of the salmon originating from the Gulf of Bothnia migrated to feed outside the gulf (Kallio-Nyberg et al. 1999). This may be the reason why the Isojoki trout need a wider feeding area. In both salmon and sea trout, stock-specifi c tendencies to migrate are maintained in foreign release sites and habitats (Hansen & Jonsson 1990, Kallio-Nyberg & Ikonen 1992, Svärdson & Fagerström 1982).

The phenotypic differences in spatial sea distribution observed between stocks could be largely genetic, because the environmental vari-ation between groups was minimized in the experiment. Moreover, the sea trout stocks stud-ied have been shown to be genetically different (Koljonen 1989). Stock-specifi c differences in some other components of the migration behav-iour of trout have been shown to have a genetic basis. For example, the genetic and stock-specifi c direction of smolt migration (Huusko et al. 1990, Koljonen & Huusko 1993) and the inherited response to the direction of the river current in the smolts of brown trout (Salmo trutta) (Jonsson et al. 1994) have been demonstrated.

The temporal sea distribution was stock dependent. The null hypothesis was thus rejected. The trout stocks were caught at different times (stock-recovery time interaction), as fi shing activ-ity varied seasonally between different sea areas or the stocks migrated seasonally between coastal and offshore waters (recovery site and recovery time interaction). The seasonal move between freshwater and the sea one or more years before

maturation is typical of the behaviour of migra-tory trout (Jonsson 1985). In winter, the Isojoki trout were mainly caught offshore. At that time, ice covers the coastal waters, where the Ingar-skilanjoki trout migrate to feed. The results do not tell us how long the trout spend in the sea. Most of them were probably caught before the fi rst spawning migration.

Effect of release site on sea migration

Both the Ingarskilanjoki trout and the Isojoki trout released into two sites on the coastline showed stock-specifi c spatial distributions inde-pendent of release site. At both sites, the Isojoki trout migrated offshore more frequently than did the Ingarskilanjoki trout. The Isojoki sea trout left the waters near the release sites, except for the release site itself, and tended to migrate to feed offshore, too. The Ingarskilanjoki trout stayed near the coastline. Release further out from the estuary, in the waters of Hernesaari, did not stimulate the Ingarskilanjoki sea trout to leave the coastal waters. The results suggest that migration distance is partly genetically and partly environmentally controlled. The attempt to reach suitable feeding areas may guide the sea migration of trout. Salmon originating from Both-nian Bay presumably remain in the Gulf of Both-nia in years with ample food resources (Salminen et al. 1994, Kallio-Nyberg et al. 1999).

Migration tendency of experimental stocks

Trout show great variation in life history, migra-tion behaviour and activity, and growth rate (Jonsson 1985), and both anadromous and resi-dent life-history forms may exist in the same population (Jonsson 1982). There is not much information on the proportions of migratory and resident forms in the natural stocks in the rivers Isojoki and Ingarskilanjoki. The reared groups at least are anadromous. Allozyme stud-ies showed that the brood stocks of the Isojoki and Ingarskilanjoki have been changed geneti-cally by breeding (Saura 1998, Ahvonen et al. 1993). Thus the experimental groups of these stocks may represent only a part of the variation


in these stocks. Trout from the two rivers have different breeding histories (Saura 1998, Kallio 1986). The Isojoki trout have been reared through many fi sh generations, since the 1960s, but the Ingarskilanjoki trout through only two genera-tions, since the 1980s. Breeding with unin-tentional or intentional selection has probably affected the genetic traits of fi sh in brood stocks as it has changed the traits of salmon (Fleming et al. 1996). Also, the differences in the ten-dency of trout to migrate have been shown to be affected by environmental selection pressure in the wild (Jonsson 1982). Large anadromous spawners were favoured when the brood stock of the Isojoki trout was established, but the brood stock of the Ingarskilanjoki trout was established with juveniles caught in the stream, without information on life history, sea migra-tion or sea growth rate. The long breeding history and the domesticating selection of the Isojoki trout may partly explain the higher sea growth rate of the Isojoki than of the Ingarski-lanjoki trout.

Both experimental stocks were established for stocking purposes. The Isojoki stock is used in its own river for improving the stock there and in other rivers for sea-ranching in the Gulf of Bothnia. Because no reared sea trout stocks originating from the Gulf of Finland existed before the Ingarskilanjoki brood stock, the Iso-joki stock has also been used for stockings in the Gulf of Finland. This comparative study showed that both experimental stocks had a tendency to migrate.

On the basis of migration behaviour, both stocks are suitable for stocking in coastal areas of the Gulf of Finland. However, the Ingarski-lanjoki trout stock is better for enhancement purposes, at least in its own river and near empty rivers, due to its genetic origin. In enhancement, the goal is to establish a natural stock with long-term natural production. The present Isojoki stock might be more suitable for sea-ranching, because its growth rate was higher than that of the Ingarskilanjoki stock after the fi rst sea year. In the future, the Ingarskilanjoki brood stock may eventually have a higher growth rate after many hatchery generations due to domestica-tion. The Ingarskilanjoki trout stock can also be recommended for sea-ranching in the Gulf of

Finland, because it originated from this area and because the strayers of the transplanted stock pose a threat to the genetic structure of the original stock (Hansen & Loescheke 1994). The management of sea trout in the future could be conducted with the aid of natural smolt produc-tion and strict fi shing regulation in some rivers and artifi cial production for fi shery in some others.

Sea migration and sea area

Both stocks were caught mainly in the coastal waters of the Gulf of Finland. The limited spa-tial distribution of recoveries observed here was similar to that reported for sea trout in other sea trout stocks (Svärdson & Fagerström 1982, Berg & Berg 1987). There are no earlier observations on the distribution of the Ingarskilanjoki trout in the sea. Tagging experiments in the Gulf of Bothnia show that the Isojoki sea trout is a stationary type in terms of spatial sea migra-tion. It was caught mainly near the release site, in coastal waters, and it migrated mainly northwards in this sea area (Ikonen & Auvinen 1984, Leskelä & Hudd 1997). The Isojoki trout released into the estuary of the Kyrönjoki in the Gulf of Bothnia were caught as adults (over 35 cm) most frequently in Finnish coastal waters, half more than and half less than 50 km from the release site. Only one fi sh was caught outside the Gulf of Bothnia (N = 318) (Leskelä & Hudd 1997). The Isojoki sea trout released into the estuary of the Kokemäenjoki in the Gulf of Bothnia also showed a tendency to migrate northwards and to remain in coastal waters (Ikonen & Auvinen 1984). Here, when Isojoki sea trout were released into the estuary of the Vantaanjoki, 11% of them left the Gulf of Fin-land, without showing, however, any clear pref-erence for direction of migration.

Sea migration and fi shery

The spatial and temporal distributions of tag recoveries also depend on the distribution of fi shery. In a comparative study, it is, however, essential that the method should not affect the dif-


ferences between groups. There was no statistical indication, that is, no stock-gear interaction, that stocks were caught with different gear types. Therefore, gear types did not infl uence the stock-specifi c differences in spatial and temporal sea distributions. Likewise, fi sh size, and release time, two factors that may affect catchability, were similar for both stocks. However, the difference in the temporal distribution of the two stocks might partly be due to the difference in the spatial distribution of the stocks in the season. The interaction between recovery site and recovery time supported this assumption. Both stocks were caught intensively in coastal waters in winter, but because the Isojoki trout is found more frequently in the open sea, they might be caught more often offshore in winter, when coastal waters are covered with ice. The Isojoki trout may also move seasonally more actively between coastal waters and the open sea.

In the management of fi sh stocks, it is impor-tant to favour measures that promote the con-servation of adaptive variation in fi sh species. Salmon populations living in their own habitat are locally adapted (Quinn 1982). In contrast, a transplanted fi sh stock has poorer fi tness, for example, homing success, than has the local population (Bams 1976). Moreover, in culti-vation the reared fi sh stock has a high prob-ability of losing its genetic variability (Verspoor 1988). Primarily, owing to the genetic threats in particular, the Ingarskilanjoki stock should be favoured in future stockings in the Gulf of Finland.

Acknowledgements

We thank Eero Jutila and Outi Heikinheimo for critically reading the manuscript and giving us many helpful com-ments. The personnel at Laukaa Fisheries Research and Aquaculture organized the rearing and tagging of the fi sh. Gillian Häkli corrected the English.

References

Anonymous 1992: Report of the Baltic and Trout Assess-ment Working Group 1992. — International Council for the Exploration of the Sea. ICES C. M. 1992/ Assess: 10. Copenhagen, 118 pp.

Anonymous 1989: SAS Institute Inc., SAS/ STAT® User’s Guide, Version 6, Fourth Edition, Volume 1. — Cary, NC: SAS Institute Inc., 1989. 943 pp.

Ahvonen, A., Jutila, E., & Koskiniemi, J. 1993: Metsäta-louden vaikutukset kalastoon Isojoen vesistön alueella: tutkimusalue ja kalaston perusselvitys. — Kalantut-kimuksia — Fiskundersökningar 69: 61–101.

Bams, R. A. 1976: Survival and propensity for homing as affected by presence or absence of locally adapted paternal genes in two transplanted populations of pink salmon (Oncorhynchus gorbuscha). — Journal of Fisheries Research Board of Canada 33: 2716–2725.

Berg, O. K. & Berg, M. 1987: Migrations of sea trout, Salmo trutta L., from the river in northern Norway. — Journal of Fish Biology 31: 113–121.

Bohlin, T., Dellefors, C. & Faremo, U. 1993: Timing of sea-run brown trout (Salmo trutta) smolt migration: effects of climatic variation.— Canadian Journal of Fisheries and Aquatic Science 50: 1132–1136.

Carlin, B. 1969: Salmon tagging experiments. — Swedish Salmon Research Institute, Report 2–4: 8–13.

Fleming, I. A., Jonsson, B., Gross, M. R. & Lamberg, A. 1996: An experimental study of the reproductive behaviour and success of farmed and wild Atlantic salmon (Salmo salar). — Journal of Applied Ecology 33: 893–905.

Hansen, L. P. 1988: Effects of Carlin tagging and fi n clipping on survival of Atlantic salmon (Salmo salar L.) released as smolts. — Aquaculture 70: 391–394.

Hansen, L. P. & Jonsson B. 1990: Restocking the river Akerselv, Oslo with Atlantic salmon smolts Salmo salar L. of different stocks. — Fauna norv. Ser. A 11: 9–15.

Hansen, L. P. & Jonsson, B. 1991a. Effect of smolt age on migratory behaviour of Baltic salmon, Salmo salar L., transplanted to the east Atlantic. — Aquaculture and Fisheries Management 22: 357–362.

Hansen, L. P. & Jonsson, B. 1991b. Evidence of a genetic component in the seasonal return pattern of Atlantic salmon, Salmo salar L. — Journal of Fish Biology 38: 251–258.

Hansen, L. P., Jonsson, N. & Jonsson, B. 1993: Oceanic migration in homing Atlantic salmon. — Animal Behaviour 45: 927–941.

Hansen, M. M. & Loeschcke, V. 1994: Effects of releasing hatchery-reared brown trout to natural trout popula-tions. — In: Loeschcke, V., Tomiuk, J. &. Jain, S. K. (eds.), Conservation genetics: 273–289. Birkhäuser Verlag, Basel.

Hindar, K., Jonsson, B., Ryman, N. & Ståhl, G. 1991:Genetic relationships among landlocked resident, and anadro-mous Brown trout, Salmo trutta L. — Heredity 66: 83–91.

Hurme, S. 1970: Lohi ja taimen Suomenlahden alueella. Maataloushallituksen kalataloudellinen tutkimustoi-misto. — Monistettuja julkaisuja N:o 37. 45 pp.

Huusko, A., Van der Meer, O. & Koljonen, M.-L. 1990: Life history patterns and genetic differences in brown


trout (Salmo trutta L.) in the Koutajoki River system. — Polskie Archiwum Hydrobiologii 37: 63–77.

Ikonen, E. & Auvinen, H. 1984: Migration of sea trout stocks in the Baltic Sea on the basis of Finnish tagging experiments. — ICES, Anadromous and Cat-adromous Fish Committee, C.M. 1984/M: 5.

Isaksson, A. & Bergman, P. K. 1978: An evaluation of two tagging methods and survival rates of different age treatment groups of hatchery-reared Atlantic salmon smolts. — Journal of Agricultural Research in Ice-land 10: 74–99.

Johannsson, M. & Einarsson, S. M. 1993: Anadromous brown trout (Salmo trutta L.) populations in southern Iceland. — ICES. Anadromous and Catadromous Fish. Committee, C.M. 1993/M: 11.

Jonsson, B. 1982: Diadromous and resident trout Salmo trutta: is their difference due to genetics? — Oikos 38: 279–300.

Jonsson, B. 1985: Life history patterns of freshwater resident and sea-run migrant brown trout in Norway. — Transactions of American Fisheries Society 114: 182–194.

Jonsson, B. & Jonsson, N. 1993: Partial migration: niche shift versus sexual maturation in fi shes. — Reviews in Fish Biology and Fisheries 3: 348–365.

Jonsson, N., Jonsson, B., Skurdal, J. & Hansen, L. P. 1994: Differential response to water current in offspring of inlet- and outlet-spawning brown trout Salmo trutta. —Journal of Fish Biology 45: 346–359.

Jutila, E., Ahvonen, A., Laamanen, M. & Koskiniemi, J. 1998: Adverse impact of forestry on fi sh and fi sheries in stream environments of the Isojoki basin, western Finland. — Boreal Environmental Research 3: 395–404.

Kallio, I. 1986: Vaelluskalakantojen nykyinen tila ja hoito. — Riista- ja kalatalouden tutkimuslaitos, Kalantut-kimusosasto. Monistettuja julkaisuja 44: 1–51.

Kallio-Nyberg, I. & Ikonen, E. 1992: Migration pattern of two salmon stocks in the Baltic Sea. — ICES Journal Marine Sciences 49: 191–198.

Kallio-Nyberg, I., Koljonen, M.-L. & Saloniemi, I. 2000: Effect of maternal and paternal line on spatial and temporal marine distribution in Atlantic salmon. — Animal Behaviour 60: 377–384.

Kallio-Nyberg, I., Peltonen, H. & Rita, H. 1999: Effects of stock-specifi c and environmental factor on the feeding migration of Atlantic salmon (Salmo salar) in the Baltic Sea. — Canadian Journal of Fisheries and Aquatic Science 56: 853–861.

Koljonen, M.-L. 1989: Uudenmaan meritaimenkantojen geneettinen tutkimus. — Suomen Kalastuslehti 3: 128–131.

Koljonen, M.-L. & Huusko, A. 1993: Genetic variation of brown trout stocks in the Koutajoki river system. — Oulanka Report 12: 129–132.

Koljonen, M.-L. & Kallio-Nyberg, I. 1991: The Finnish

trout (Salmo trutta) stock register. — Finnish Fisher-ies Research 12: 83–90.

L’Abee-Lund, J. H., Jonsson, B. Jensen, A. J., Saettem, L. M., Heggberget, T. G. Johnsen, B. O. & Naesje, T. F. 1989: Latitudinal variation in life-history characteris-tics of sea-run migrant brown trout Salmo trutta. — Journal of Animal Ecology 58: 525–542.

Marttinen, M. & Koljonen, M.-L. 1989: Uudenmaan mer-itaimenkantojen investointi ja geneettinen tutkimus. — Uudenmaan kalastuspiirin kalastustoimisto. Tie-dotus nro 3. 141 pp.

McCullagh, P. & Nelder, J. A. 1989: Generalized linear models. Monographs on statistics applied probability. — Chapman and Hall. London. 511 pp.

Pratten, D. J. & Shearer, W. M. 1983: The migrations of North Esk sea trout. — Fish. Mgmt. 14: 1–9.

Quinn, T. P. 1982: Intra-specifi c differences in sockeye salmon fry compass orientation. — In: Brannon, E. L. & Salo, E. O. (eds.), Proceedings of the Salmon and Trout Migratory Behaviour Symposium: 79–85. University of Washington Press, Seattle, WA.

Salminen, M., Kuikka, S., & Erkamo, E. 1994: The divergence in the feeding migration of Baltic salmon (Salmo salar L.); the signifi cance of smolt size. — Nordic Journal of Freshwater Research 69: 32–42.

Salminen, M., Kuikka, S., Erkamo, E. 1995: Annual variability in survival of sea-ranched Baltic salmon, Salmo salar L.: signifi cance of smolt size and marine conditions. — Fisheries Management and Ecology 2: 171–184.

Saura, A. 1998: Suomenlahden meritaimen. Kalastuksen ja hoidon kehittämissuunnitelma. — Kala- ja riista-raportteja nro 110. 22 pp.

Skaala, O. & Naevdal, G. 1989: Genetic differentiation between freshwater and resident and anadromous brown trout, Salmo trutta, within watercourses. — Journal of Fish Biology 34: 597–605.

Skrochowska, S. 1969: Migrations of the sea trout (Salmo trutta L.), brown trout (Salmo trutta m. fario L.) and their crosses. Part III. Migrations to, in and from the sea. — Polish Archives of Hydrobiology 16: 149–180.

Sturlaugsson, J. & Johannsson, M. 1996: Migratory pat-tern of wild sea trout (Salmo trutta L.) in SE-Iceland recorded by data storage tags. — International Coun-cil for Exploiration of the Sea, Anadromous and Catadromous Fish Committee, C.M. 1996/ M: 5.

Svärdson, G. & Fagerström, Å. 1982: Adaptive Dif-ferences in the Long-Distance Migration of Some Trout (Salmo trutta L.) Stocks. — Rep. Inst. Fresh. Research, Drottningholm 60: 51–80.

Toivonen, J. & Ikonen, E. 1978: Havsöringen i Finland. — Fiskeritidskrift för Finland 5: 104–109.

Verspoor, E. 1988: Reduced genetic variability in fi rst-generation hatchery populations of Atlantic salmon (Salmo salar). — Canadian Journal of Fisheries and Aquatic Science 45: 1686–1690.

Documents

Sea migration pattern of two sea trout (Salmo trutta ... · Zool. Fennici 39: 221–235 ISSN 0003-455X • Sea migration of two sea trout stocks 221 ... Sea migration pattern of two